I don't think 2 years is typical for a PhD (even after already earning a M.Sc)

This may be just a sign of the quality of research Sergey was doing which has, as we know now, fairly useful applications :-)